IRC channel logs

2019-08-22.log

back to list of logs

<ngz>minall: You'll probably learn a lot doing that.
<minall>ngz: Yes, I'll probably do it since I'll learn a lot (so fun)
<minall>But I don't thing I'll do it soon, but I'll try
<Sharlatan>Hi, how to trouble shoot "hash mismatch" issue for `git-fetch` ?
<rekado>Sharlatan: get the hash from a known good checkout with ‘guix hash -rx’
<rekado>I assume that this is for a package you are working on?
<Sharlatan>yes, rekado. trying to package `alacritty` terminal (Rust based)
<Sharlatan>guix download https://github.com/jwilm/alacritty/tree/v0.3.3
<Sharlatan>gives me hash `1zgl5gndaxwxknc6cl4kh8wqggqfxzh0gp2s4kbq2sh3bz39hn7i`
<rekado>Sharlatan: ‘guix download’ doesn’t use git-fetch
<rekado>you’re fetching a website.
<Sharlatan>ok, let me check rekado
<rekado>use ‘git fetch https://github.com/jwilm/alacritty.git’, checkout ’v0.3.3’, and then use ‘guix hash -rx’ to get the correct hash.
<Sharlatan>rekardo I could manage to path hash issue, now guix could find `clap` package in `https://github.com/rust-lang/crates.io-index`, I've checked it manually and it's there.
<quiliro>‘git fetch https://github.com/jwilm/alacritty.git’ does not work
<rekado>sorry, ‘git clone’
*rekado –> zzzZ
<adom`>Hi! I'm sorry to just barge in, but is this an okay place to ask for help? I've been reading tutorials and mailing list threads but I'm totally stuck.
<minall>adom`: You can ask for help just fine!, we are here for that actually
<adom`>Okay, well hopefully I won't take too much of your time. So, the problem is that when I run guix pull, I get the following output:
<adom`>guix pull: error: symlink: File exists: "/var/guix/profiles/per-user/adom/current-guix"
<adom`>Also, ~/.config/guix is totally empty. Although there is stuff in /root/.config/guix
<adom`>I can install packages as a user, so I'm not sure if it's even a problem that there's nothing in ~/.config/guix.
<adom`>Guix describe says that it "failed to determine origin"
<minall>Is that a recent reconfigure?
<adom`>I've just installed for the first time. I tried once by following the instructions in the installation chapter, and then uninstalled and used the script. Same error both times.
<adom`>Is it better to keep posting stuff about my config, or to let you ask questions that you deem relevant?
<adom`>(Thanks a bunch!)
<quiliro>adom`: did you install guix pcakage manager on a foreign distro or the guix system distro?
<quiliro>package*
<adom`>Foreign distro (ubuntu)
***jonsger1 is now known as jonsger
<quiliro>i am not sure about your problem with users
<quiliro>each user has their own profile which points to its configuration of its guix
<quiliro>maybe if you try another user you can compare
<adom`>Like, I should try to run it with one of the fake users it created? (I don't have any other accounts on the computer)
<adom`>There should be something in ~/.config/guix, yeah?
<adom`>Yeah, when I looked at the profile for "adom" it was entirely commented out. The root profile was the same.
<adom`>guix pull -l outputs:
<adom`>guix pull: error: profile '/var/guix/profiles/per-user/adom/current-guix' does not exist
<quiliro>adom`: nt the guix users..your users
<quiliro>*not
<quiliro>adom`: i have seen that error on the mailing list archives
<quiliro>adom`: did you check for that error?
<quiliro>i cannot find it in help-guix, will search on bug-guix
<adom`>I'll look a bit more for that error. I was mostly searching for a different one. I may have to reinstall though
<adom`> https://www.mail-archive.com/help-guix@gnu.org/msg04842.html
<adom`>The person there seemed to have a similar problem and they fixed with `rm current-guix-1-link`
<quiliro>i see that....try it!
<adom`>but now that I ran that I can't use guix anymore...
<quiliro>:-(
<adom`>I don't have /run/current-system
<adom`>do you?
<quiliro>i use guix system...let me see
<adom`>Thanks!
<adom`>And no need to stick around here all night. I'm sure I'll figure something out eventually. Or come back later :)
<quiliro>$ ls /run/current-system
<quiliro>boot etc initrd kernel locale parameters profile
<quiliro>
<quiliro>adom`: if i don't abandon, you don't abandon! ;-)
<adom`>and then do you have profile/bin/guix?
<adom`>it's weird that I don't have that
<quiliro>adom`: under which directory?
<adom`>under /run/current-system
<quiliro>$ ls -l /run/current-system/profile/bin/guix
<quiliro>lrwxrwxrwx 7 root root 73 dic 31 1969 /run/current-system/profile/bin/guix -> /gnu/store/70d4vzl7mk5r3n0yn8ah94xxmsn9la4r-guix-1.0.1-3.4a54ed7/bin/guix
<quiliro>so it does exist
<adom`>hmm I don't know why I'm missing all of these things
<quiliro>adom`: did you enter another user and test 'guix pull'?
<adom`>I'm going to reinstall, and then I can try that
<quiliro>maybe you were messing with files and erased things
<adom`>because now I can't figure out how to do anything
<adom`>but I'll get back to you when I do!
<quiliro>ok...but you will learn more if you repair
<quiliro>adom`: even if it does not work, finding the structure of things is a great learning experience
***card.freenode.net sets mode: +o Sigyn
***grubles is now known as Guest48292
***jamesrichardson is now known as Guest91833
***Guest64341 is now known as nalkri
<jgibbons[m]>I have two drives in my computer. I want to use one as my root and one as my /home. ext4 doesn't like links between separate drives. Will this cause trouble with my profile ($HOME/.guix_profile or something like that)? If so is there a known workaround?
<erudition1>I've never heard of that issue with ext4, and I don't know the answer, but I do know that btrfs is supported and works well, with a lot more useful features
<erudition1>Less of a workaround, more like making the problem irrelevant heh
<efraim>As far as I know it won't be a problem
<str1ngs>jgibbons[m]: I have a /home and / different partitions it's not a problem
<alloy>Hi! I've setup offloading to a state where guix offload test works. But how do I enable distributed builds on a normal guix build run? Thanks!
<nckx>alloy: What does Guix say?
<alloy>When I pass -M 0 it says it can't start any build and I should either increase max-jobs or enable dirstributed builds
<nckx>Heh, when I pass -M0 it builds locally anyway 😛
<nckx>(And I have offloading set up.)
<alloy>Okay, strange
<nckx>When running any Guix command that builds anything, Guix will connect to all remote nodes and should print something about their ‘normalised load’ (which is the remote load divided by that machine's parallel-builds). If it's over 2 the machine will not be used. Are all your machines highly loaded?
<alloy>I've got a normalized load of 0.08, doesn't seem to overloaded :D
<alloy>Got it from guix offload status
<alloy>But it won't print that on a normal guix build run for me
<nckx>Strange. ‘guix build <random stuff>’ gives me http://paste.debian.net/1096947/
<nckx>I never had to do anything besides write machines.scm and get ‘guix offload test’ to work.
<nckx>I don't think there's a separate ‘now use it’ knob.
<alloy>Mhm, okay. Thanks anyways!
<nckx>Wish I could be of more help. Nothing offloading-related at all in your ‘guix build’ output?
<alloy>Nope
<civodul>Hello Guix!
<efraim>sneek: botsnack?
<sneek>:)
<efraim>sneek: later tell ng0 I have inconsistent results with some crates using 'extern crate' on some deps and not others, and shared deps freaking out the compiler
<sneek>Got it.
<efraim>sneek: later tell ng0 libgit2-sys can almost link to libz which can link to libc, but libgit2-sys doesn't like that libz isn't linked to the vendored libc source it's using, so that's where I'm stuck for now
<sneek>Got it.
<efraim>sneek: botsnack
<sneek>:)
<efraim>TIL 'git stash save' is depreciated and 'git stash push -p' is AMAZING
<civodul>"git stash push"? woow
<civodul>-p looks nice indeed
*civodul hopes Magit supports it
<efraim>I also just learned that as part of cargo's 'cargo build' command it creates a Cargo.lock file, so we can get rid of the 'update-cargo-lock phase from the cargo-build-system and move the deletion of Cargo.lock to another phase
***Guest48292 is now known as grubles
<civodul>efraim: oh, you should ping Ivan, Robert, and all on the list
<nixo_>rekado_: hi, thanks for the guix-artwork merge! Do you know how frequently the website is generated (i.e., when the packages.json will be available)?
<civodul>hey nixo_!
<civodul>nixo_: it's rebuild every hour
<civodul>*rebuilt
<nixo_>hi civodul! Thanks, then maybe something is not working (or am I checking the wrong url)? I'd expect after 9c6f714305460e99c681d9b7f368e13bfe49fdd9 was merged to see this page: https://guix.gnu.org/packages.json
<dfeng>Hi!
<nixo_>also, the number of packages in my guix is greater (10424) than the number reported https://guix.gnu.org/packages/ (9,789)
<civodul>nixo_: when you "guix build -f build.scm" on guix-artwork/website, do you get packages.json?
<dfeng>I have not read the manual yet. I would like to know if there is a kind of sandbox to experiment with package definition?
<civodul>dfeng: package builds always happen in a "sandbox" (chroot + separate user namespaces)
<dfeng>civodul: I also mean a temporary space which can be erased.
<civodul>dfeng: the build tree is automatically erased, and the build output ends up in /gnu/store, where it is subject to garbage collection
<jonsger>nixo_: the number at https://guix.gnu.org/packages/ is the number at version 1.0.1, not on master
<dfeng>civodul: okay, thank.
<civodul>see https://guix.gnu.org/manual/en/html_node/Features.html
<nixo_>civodul: There's no "build.scm" file in there, but on the readme the way to run it is "GUIX_WEB_SITE_LOCAL=yes haunt build". Then yes, with haunt serve I have the json file
<nixo_>jonsger: Ok thanks, that explains the difference.
<civodul>nixo_: sorry, it's .guix.scm
<civodul>nixo_: re stale packages, see https://issues.guix.gnu.org/issue/36708
<nixo_>civodul: the build is failing :|
<nixo_>error: guix-url: unbound variable
<nixo_>but /tmp/gnu.org/software/guix/packages.json exists
<chrislck>further noob questions: guix install allegro --> long compile; guix remove allegro; guix gc; guix install allegro --> quick. does this mean allegro compiled to my cpu is now cached on ci.guix.gnu.org?
<efraim>it means that there wasn't a substitute available on ci.guix.gnu.org so you built it locally. when you removed it and re-installed it, it was already cached in your /gnu/store directory.
<efraim>all the packages on ci.guix.gnu.org are built on the build farm behind ci.guix.gnu.org
<chrislck>so it means /gnu/store will grow steadily with time, unchecked
<efraim>until you run 'guix gc', to remove the unreferenced packages
<chrislck>well see my post above; guix gc should remove the built allegro shouldn't it?
<efraim>I normally use 'guix gc -F 50G' to remove packages until I have about 50GB of space available on my drive
<efraim>the above allegro is likely still referenced by a previous profile generation so it won't be removed yet
<chrislck>gah
<chrislck>but if the allegro was referenced by previous profile, then it shouldnt need to be compiled the first time i install it?
<efraim>it's part of the atomic package management, so you can roll-back to a previous state if you want to
<chrislck>learning curve is too steep
<chrislck>too much smoke and mirrors
<efraim>it only needed to be compiled because there wasn't already a binary substitute from the substitute server, which happens sometimes
<kmicu>chrislck: without seeing logs it’s guessing what happend. It’s possible that allegro showed up on a substitute server before the second ‘guix install allegro’ was executed.
<divansan`>I'm launching chromium like so, CHROMIUM_ENABLE_WEB_STORE=1 /run/current-system/profile/bin/chromium but still when I browse to an extension on like https://chrome.google.com/webstore/detail/revolver-tabs/dlknooajieciikpedpldejhhijacnbda?hl=en it doesnt have the add to chrome button.
<divansan`>perhaps that ability has been lost since it moved into guix codebase...
<civodul>nixo_: did you find out where that unbound variable comes from?
<nixo_>civodul: no way, the backtrace is a mess. And going back with git I have even more errors, so I couln't even bisect
<civodul>uh
<civodul>nixo_: "guix build -f .guix.scm" works for me as of Guix 178f1d1f75a8d512b517f474f63da507beaebd49 and artwork.git aedd60ae9b05734ef85a654c8f3fea61f009eebd
<nixo_>my guix version is 524c0ece8ec660afc3675767c96b31e516158ea8
<civodul>trying with a newer artwork.git now...
<kmicu>divansan`: I assume ungoogled‑chromium works that way https://ungoogled-software.github.io/ungoogled-chromium-wiki/faq#can-i-install-extensions-or-themes-from-the-chrome-webstore
<civodul>nixo_: also works with artwork.git 9c6f714305460e99c681d9b7f368e13bfe49fdd9, and this time there's packages.json
<civodul>so why is it not on-line, hmm...
<civodul>ooh, i know: the Guix used to build the web site on berlin still uses Guile-JSON 1.x
<civodul>interesting
<quiliro>Saluton Gixujo!
<quiliro>*Giksujo
<nly>hello
<minall>Hello guix!~
<minall>What are the steps to package kde? any idead?
<nly>i suppose there would be several sub packages, some core ones, and some peripheral ones
<nly>there might have been some previous work on kde which might help
<nixo_>civodul: I'm on a clean checkout of guix-artwork 9c6f714, guix --version = f5111 (slightly newer than yours). But still guix build -f .guix.scm does not work. Did on 2 different machines (other has guix version 3251c628f754). Is it "cd guix-artwork/website; guix build -f .guix.scm" ? I don't get why our results differ
<civodul>nixo_: that's because ".guix.scm" refers to the 'guix' package, and presumably yours uses guile-json@1 whereas mine uses guile-json@3
<civodul>see "guix describe"
<civodul>and "guix package --show=guix"
*civodul tries to see how we can improve on that
<xavierm02>Hey. So by removing the texlive-union from the propagated inputs of the lyx package, everything seems to work fine (i.e. if you don't have texlive / texlive-bin lyx runs but when you try to export to pdf there's an error, and if you have them everything works)
<xavierm02>The problem is: I don't know why this texlive-union was ever added...
<nixo_>civodul: ok I can confirm that my guix package --show=guix is at version c902458863, the commit before the move to json3 (2eb0628a42)
<civodul>xavierm02: hi! texlive is probably required by lyx, but the problem might be that it shouldn't be propagated, no?
<xavierm02>civodul: Well to edit files, you don't need it. I'd think of texlive as "recommended" for lyx, but I don't think there's "recommended" in guix
<xavierm02>civodul: Also, the reason I want to remove the texlive-union is that if it's here, LyX uses it, so that even if you install the full texlive, it'll only find stuff from the texlive-union
<l3nn>hi, is anybody visiting cccamp 2019?
<jlicht>hey guix!
<civodul>xavierm02: indeed, good points; i'd suggest asking on guix-devel, Cc'ing people who touched it
<civodul>hey jlicht!
<civodul>l3nn: dunno!
<jlicht>hey civodul, welcome back! haven't seen you online for a while
<civodul>yup, i was happily offline, and i'm happy to be back here now! :-)
<civodul>nixo_: if you pull from artwork.git and try again, it should be better
<adom`>Hi, I was in here last night but I was wondering if I could bother anyone here for help getting guix installed.
<adom`>On a fresh install, when I first run guix pull I get:
<adom`>Migrating profile generations to '/var/guix/profiles/per-user/adom'...
<adom`>guix pull: error: while creating symlink '/home/adom/.config/guix/current': Permission denied
<civodul>hi adom`1
<civodul>is it on Guix System or on another distro?
<nckx>I got that a few times on a few systems, removing that directory and running ‘guix pull’ again fixed it.
***dongcarl_ is now known as dongcarl
<nckx>You'll have to ’readlink -m $(which guix)’ before that to print the full path of the guix you'll have to run.
<adom`>Hi! I'm installing on a foreign distro. Thanks nckx. I'll try that!
<rekado_>xavierm02: I think I added it back then. That was part of the move from the monolithic texlive package to the much smaller texlive-union. I remember that TeX Live was required at build time. Perhaps propagation is wrong here.
<rekado_>I’m still working on a fix for our modular TeX Live (found another problem with our font maps)
<rekado_>I’ll create a new branch for the fixes soon.
<rekado_>there’s just one more odd problem now that the font maps are fixed (it doesn’t find any Latex files any more…)
<rekado_>civodul: welcome back!
<adom`>Nckx: To clarify in case I'm about to mess things up. Now that I have the path /gnu/store/.../bin/guix, I rm -r ~/.config/guix/current? Or the whole ~/.config/guix?
*rekado_ –> afk
<civodul>hey rekado_!
<adom`>nckx: When I delete that, I get the following error: guix pull: error: symlink: File exists: "/var/guix/profiles/per-user/adom/current-guix".
<minall>Why does guix uses herd? instead of systemd, is there a reason for this?
<str1ngs>minall: shepherd makes perfect sense, its extensible in scheme. so interaction between guix services and shepherd is trivial
<nixo_>civodul: thanks, but still the same. There's the long "Updating channel" phase before, but then it throws the same error (guix-url). Now I need to go, I'll have a look again later, thanks!
<minall>str1ngs: Thank you, makes sense
<adom`>When I delete /var/guix/../adom/current-guix, I then get this error.
<adom`>Updating channel 'guix' from Git repository at 'https://git.savannah.gnu.org/git/guix.git'...
<adom`>guix pull: error: mkdir: Permission denied
<adom`>It seems like there's something wrong with the permissons for guix? Is there something I can do to check that directly? Or change it? Or a stage in the installation process I should take a closer look at?
<nckx>adom`: No explanation of *what* it was trying to mkdir? :-/
<adom`>Oh sorry. It has this first: Migrating profile generations to '/var/guix/profiles/per-user/adom'... Nothing else though.
<nckx>adom`: /var/guix/profiles/per-user/adom should be owned by adom:adom.
<nckx>Is that the case?
<nckx>(Looking at my own Guix System here.)
<adom`>Yeah, owned by adom:adom. (Sorry for all of the trouble -.-)
<nckx>No trouble, I just don't understand which mkdir is failing then.
*civodul pushes a bunch of 3-week-old patches
<adom`>Alright! I ran again with sudo, then deleted .../per-user/root/current-guix, then ran again with sudo... And something is actually happening!
<nckx>Huzzah!
<adom`>We'll see what happens :)
<adom`>I'm getting warnings that it can't set locale. Not sure if that's bad.
<nckx>Those are harmless.
<jlicht>adom`: you are running guix on top of another distro, right?
<nckx>Yes.
<adom`>jlicht: Yeah ubuntu 18.04.01
<jlicht>it should be okay then, after you get `guix pull' to work, https://guix.gnu.org/manual/en/html_node/Application-Setup.html has some instructions on fixing that :-)
<jlicht>annoying thing is that you have to do it for all guix profiles (so in your case both root and adom)
<adom`>Even if it's the root version of guix pull? I'm worried that it won't work for adom after this.
<jlicht>I have never had good experiences with having a separate root profile when using guix, so I can't help you there
<adom`>Cool cool. I'll check out the link after this finishes. Thanks :)
<adom`>Man. IRC is really so amazing. I need to find some channels where I can be helpful.
<nckx>adom`: I'm not a fan of running ‘sudo guix pull’ either (because in 99% of cases it doesn't do what people thought it was for), but if it gets you a working Guix now I'd say that's a win & we can fix the permissions later.
<nly>i don't think i've run 'sudo guix pull' in months, i just use the user profile
<nly>'sudo -E guix something' when you need permission, only reconfigure needs that in my usage
<nckx>nly: I've never run it, period. Why did you? Just curious what its use cases are.
<minall>quiliro: Kiel vi fartas!
<adom`>Okay it just finished running. It says that "The following env variable definitions may be needed: .." Do I just run the export path="whatever" or do I put that somewhere?
<adom`>(In my bash profile?)
<adom`>I got excited and tried to run guix pull again, but I get the same error as before.
<nckx>adom`: They are already in your bash profile. Opening a new terminal should give you a shell with them set.
<nckx>adom`: Could you paste the output of the failing pull to paste.debian.net?
<nckx>I'm not very hopeful, but maybe there's the name of a directory in there somewhere.
<adom`>Not much there unfortunately: http://paste.debian.net/1096987/
<adom`>Can I get it to print more? Should I just use sudo guix forever...?
<nckx>I'd advise against using ‘sudo guix’ at all, in general.
<nly>nckx, i didn't know 'sudo -E' before haha so had to pull as root then update
<adom`>Yeah I guess that seems to be the consensus.
<nckx>adom`: So ‘guix=$(readlink -m $(which guix)); rm /var/guix/profiles/per-user/adom/current-guix; $guix pull’ or so doesn't work?
<nly>not 'had to', but that's the thing i knew at the time
<nckx>nly: Thanks 🙂 ‘sudo’ on Guix System is equivalent to ‘sudo -E’ on some other distributions, which doesn't help the confusion, but *is* convenient for us System users.
<adom`>Unfortunately no. More permission denied errors: http://paste.debian.net/1096988/
<nckx>adom`: Ah, at least that gives us a clue! Seems like something (maybe ‘sudo guix’) created your user's ~/.cache/guix as root. Since it's a cache, you can just ‘sudo rm -rf $HOME/.cache/guix’ and try again.
<nckx>It should at the very least print a different new & exciting error message, or you might even get lucky.
<adom`>I may have messed things up. After using that guix variable once, it doesn't seem like I have access to it anymore.
<nckx>adom`: What is it?
<nckx>adom`: It should point to the guix in the store and be independent of any symlink magic that might've been destroyed. So it should™ really work.
<adom`>Hmm, I was mistaken. You're right
<adom`>And oh my gosh
<adom`>I'm getting the same "something is happening" thing as before. Just a laggy terminal so far
<adom`>I'm embarrassed. I probably should have said earlier that I've installed and uninstalled guix a few times today and yesterday trying to get this to work. I followed the directions on the Arch wiki and I thought I was removing every trace. But I guess there was stuff in the cache.
<adom`>Wow, it really seems to be working. Incredible.
<adom`>Thanks! (fingers crossed)
<lfam>Glad it worked :)
<xavierm02>I submitted my patch!
<nckx>adom`: (cautious) \o/ !
<nckx>If it does, run a plain ‘guix pull’ (no $) just to make sure all new paths are properly set up.
<adom`>What's the appropriate thank you in situations like this? Donate to guix? Some charity you like? I really feel like I could have struggled with this for days and maybe would have just given up. I read so many email threads and irc logs and didn't see anything about the cache.
<nckx>adom`: That would be very generous of you 🙂 There's a page at http://guix.gnu.org/donate/.
<adom`>Will do (although can't give too much atm). Okay, so now it says "after setting 'PATH', run hash guix
<adom`>Is setting path just that export command from before?
<nckx>There's also Guix Europe which handles payments in euros (so we don't have to futz with exchange rates as much), but I'm not sure whether that accepts direct donations.
<nckx>And I really should know this, sorry. Just got back from holiday is my excuse 🙂
<nckx>adom`: That should do it, as will opening a new shell/terminal.
<nckx>(In the latter case no ‘hash guix’ is needed.)
<adom`>Ahh cool. Ran that in a new shell and was worried about not seeing an output. Okay I ran guix pull one more time in a new shell and it said nothing needed to be done! Wow so I think we're good?
<adom`>Is there anything else I should do to check?
<nckx>adom`: Sounds like you're all set.
<adom`>I'll donate in BTC so shouldn't be any problem with transaction fees
<nckx>adom`: Thanks for your generosity!
<adom`>Just sent $10. Thank you so much for your help. And patience. And everyone else too.
<dwdv>Hello everyone! Say, except for non-free firmware /wireless, what else would I be missing when switching over to guixsd from fedora? ffmpeg and such seems to by included.
<dwdv>Also, is guix supporting something similar to freebsb's port flavours, where a pkg has different build options (e.g. no perl support in urxvt), so multiple binaries are build, not just the main one?
<dwdv>Are vim and vim-full the exception or does this happen more often?
<nckx>dwdv: Missing: hard to tell, Guix has just over 10k packages but it's still easy to find that one killer app that we're missing.
<nckx>dwdv: Flavours: Some packages have a -full or more often a -minimal variant, but those tend to be added in an ad-hoc manner when someone needs them (e.g. emacs-no-x), not systematically.
<nckx>I'm sure we have far less of those than FreeBSD.
<jlicht>dwdv: if you know some guile, you can rewrite parts of the dependency graph of a package programmatically, (which is usually how we support python2 variations of python packages, if you are interested in an example), but this is by no means as 'easy' as using ports.
***psachin is now known as psachin|away
<minall>I need to make a reconfigure of my GuixSystem installation, but I can only do it from the live-USB
<quiliro>if the efi partition was not mounted, on boot GS will not boot...what must i do if i have this problem?
***len1 is now known as lenn
<minall>How should I do it? I mounted my installation on /mnt, and then my boot device on /mnt/boot, does chroot /mnt and -guix system reconfigure- works?
<quiliro>i am helping minall...my problem is minall's problem
<minall>I am your problem???! JK
<quiliro>no!
<minall>Just kidding lol
<quiliro>i have the same problem because your problem is my problem too....i want help. ;-)
<quiliro>did you check the mailing list? minall
<minall>On it
<quiliro>minall: maybe this works or gives you pointers to follow https://www.unixmen.com/fixing-a-few-common-grub-errors-broken-bootloader-and-error-1715/
<quiliro>but after that you should repair your config.scm and reconfigure
<quiliro>instead of update-grub
<quiliro>this is funny...not very good quality though:
<quiliro>mpv https://liberty.sfo2.digitaloceanspaces.com/see-your-junk.mp4
<quiliro>
<minall>jaaja
<minall>nice vide!
<quiliro>minall: (hd0,1) is first hard dist, first partition: sda1
<minall>i didn't knowabout librem, ill check it
<quiliro>later
<quiliro>dont have time
<quiliro>did you solve your grub problem?
<quiliro>i guess you do not have any more time
<minall>Nope,but i want to fix it
<quiliro>maybe if you check the grub manual
<quiliro>or a new tutorial
<quiliro>i do not know if that link i gave you is updated
<quiliro>minall: ^
<dwdv>nckx, jlicht: thanks for the info, will make sure to read the packaging tutorial. Regarding missing packages: not so much about the obscure ones that you use once a leap year, more about hardware acceleration support and stuff that's impacting the desktop experience.
<dwdv>Guess I'll just give it a spin, thanks again.
<bandali>hi guix, does anyone have a guix channel for using org master?
<rschifflin>hey channel, i just received a pinebook so i can finally build myself a guixbox =)
<rschifflin>whats cool is that amazon even provides arm-based ec2 instances now so i can build an installation image in a reasonable amount of time
<rschifflin>i have hit a problem though- apparently the linux-libre kernel that gets compiled when specifying `pinebook-installation-os` is erroring out trying to make some arm-neon intrinsic with some type alias collision. Here's the actual log output: https://pastebin.com/drKPqBAa
<rschifflin>anyone have any experience building guix for arm? would it be straightforward to modify my installation file to set `CONFIG_KERNEL_MODE_NEON` to `n` as a bandaid to build without the conflicting files
<Tirifto>Hello all! Saluton al chiuj!
<rschifflin>o/
<nothingmuch>is it possible to avoid generating manpage db, mime db etc when using guix environment?
<pinoaffe>is there a way to estimate the remaining build time for a package?
<pinoaffe>qt has been goin' strong for ~2 hours
<bavier>pinoaffe: nothing great for that, besides previous personal experience. You could check the build farm logs for how long the same build has taken on those machines
<bavier>I had dreamed at some point of collecting build stats and creating a model so that we could display an estimate
<bavier>some packages will show a progress bar if that information is available from the build system.
<joshuaBPMan>hello, I am currently trying to install guix, but the graphical install is not working. I know how to install manually, which is what I've done before...
<joshuaBPMan>the last couple of times I ran the manual install, half-way through the install I got a "general fault" error.
<joshuaBPMan>I googled it, and per Intel it has something to do with accessing memory that you are not supposed to.
<joshuaBPMan>I'm wondering if the graphical install is trying to run X, and it is just consuming memory until it runs out of memory.
<joshuaBPMan>is there a way to halt any graphical part of the install?
<joshuaBPMan>herd stop <service name>
<joshuaBPMan>herd disable <service name>
<joshuaBPMan>anyway, I'll give that a try.
<joshuaBPMan>actually now that I think about it, I have not have started the hurd start cow-store....
<lfam>s/hurd/herd
<joshuaBPMan>yup
<joshuaBPMan>hmmm. I'm not seeing a xorg or wayland service....I'm not certain what the service name would be...
<joshuaBPMan>I could try herd stop console-font-tty1
<joshuaBPMan>also it seems like guix system init /mnt is atomic... How is that?
<joshuaBPMan>ie: it succeeds or it fails. No inbetween. Updates being atomic makes sense. It's just messing with symlinks.
<joshuaBPMan>but with init you are copying files....
<civodul>joshuaBPMan: well the last bit is atomic
<civodul>but in this case it really depends on what you mean by "atomic"
<civodul>it definitely modifies the contents of /mnt in a non-atomic way :-)
<joshuaBPMan>well, my point is changing a symlic is super fast.
<civodul>and atomic
<civodul>yes
<joshuaBPMan>copying 1GB+ of files is not atomic. It takes several seconds to happen. and/or minutes to copy.
<civodul>yes (though atomicity and speed are two different things)
<joshuaBPMan>civodul: are they? Suppose you are copying the files. Midway through copying, your computer loses power and shuts off. That's a half finished state right?
<bgardner>Hello guix, does anyone have any sample configs for a basic samba server?
<str1ngs>joshuaBPMan: as civodul mentions. atomic also means it completes fully or not at all. it does not mean instantaneous
<quiliro>what bgardner asks for on samba, would be nice for several configurations...one i would like is guix with email server justa edit user names, domain name and ip address...
<quiliro>a mail server with spamassassin and spf and letsencrypt and clamav all in just a file
<bgardner>Maybe a little pool of example config files. Seems like it would help a lot of people.
<ngz>That's an interesting idea.
<ngz>Maybe the best way to move forward is to start writing such examples, even if they are inaccurate or plain wrong, in a wiki fashion, and improve them later.
<ngz>I mean, it is simpler for a knowledgeable person to fix an existing template than to write everything right from scratch.
<bandali>ngz, i’m curious to hear about your setup as an org dev. do you redefine emacs-org in a separate channel to fetch from latest master? or perhaps simly clone into a folder and add it to your load-path?
<bandali>*simply
<ngz>The latter.
<bandali>cool, thanks
<ngz>I don't even use channels. I'm old-fashioned, I still set GUIX_PACKAGE_PATH
<bandali>ah :)
<bandali>channels have been working pretty great for me so far
<ngz>Channels sound nice. But the environment variable gets the job done, so... Laziness, etc. :)
<bendersteed>ngz: well channels are really low maintenance in fact, if you write the package definitions a channel is five minutes away
<bandali>ngz, i understand :)
<bandali>and +1 bendersteed
<ngz>I spend my time in a much more useful way, e.g., trying to package frozen-bubble game ;)
<bendersteed>hahah of course! for me my channel is my testing ground before I commit a package
<bandali>hehe
<len1>Hi, I am wondering how to include go as a dependency for a package definition?
***len1 is now known as lenn
<bendersteed>you add it at the appropriate input group in your package definition
<bendersteed>oh this reminds me that I want to package hugo
<lenn>I've came this far: https://pastebin.com/r6kHRt9W and I am getting error that there is no code for module go
<bendersteed>is the go toolchain complicated? i've never got time to play with this
<lenn>(when I try to build ddev package)
<bendersteed>it's (gnu packages golang)
<lenn>oh
<lenn>let me try
<lenn>bendersteed: ty it solved that issue.
<bendersteed>:-)
<lenn>now I have a problem with docker-compose
<bendersteed>the docker-compose package is defined in the (gnu packages docker) module
<lenn>yes, figured it out :)
<lenn>tnx
<bendersteed>check them out, if you use emacs then emacs-guix is nice to search for info or packages
<bendersteed>should be doable from the guix command as well, but I don't use it that much so I don't know
<lenn>Yeah, I am investing my next hour in learning emacs-guix
<bendersteed>whoa, I need to package around 45 libraries for hugo
<bendersteed>what kind of sorcery is this
<lenn>wow